Replies

Agreed, and my primary point in all this, but I don't see why my post confused you. You keep asserting that the Reps are oversampled. Just because they used more Reps than Dems does not necessarily mean Reps were oversampled. It's good that you like facts, so let me try to explain it with a specific example, as opposed to a general thesis. Suppose Reps lie about their choice more often than Dems - JUST SUPPOSE - and suppose that Newsweek KNOWS this from historical data. Lets also suppose Newsweek KNOWS what value to ascribe to this dishonesty phenomenon. Then it is NOT the case that they have oversampled JUST because they called more Reps. Remember, that's just one example. There could be any number of factors like that. But the point is that YOU don't know what they are, Newsweek does. So you can't legitimately do what you're trying to do, that is, make assumptions without all the relevant information. By delving into the internals of something you really know nothing about you're just tripping over yourself.

I don't agree with your premise here at all - First off, the fact is NewsWeek has a "history" that needs to be looked at - You cannot just look at the this Newsweek poll in a vacuum - Historically NewsWeek polls are junk - They jump all over the place for whatever purpose NewsWeek wants them to serve -

Secondly I would offer you some words of wisdom from Thomas Sowell (brilliant man) - "there is nothing more complex then avoiding the obvious" -

And that is exactly what you are trying to do here - The obvious fact is Republicans we're well oversampled and Dem's were well UNdersampled in the latest NewsWeek poll - This is what clearly brought about the 11pt lead for GWB! -

If a more accurate weigthing was applied GWB would not have an 11pt lead - It would be closer to 5pt - (that is the facts) -

I don't care what agenda or purpose NewsWeek wants to have for under or over sampling - facts remain facts -

ANd lastly we need to wait until both IBD and Gallup come out with polls to see where this race stands.

And lastly we need to wait until both IBD and Gallup come out with polls to see where this race stands

We're in violent agreement on that, as we both already posted. As for your argument, you're just going in circles. Your basic theme is that facts don't matter so you'll just assume whatever you want to get the result you want. I see an agenda here.
Thanks for your reply, your logic is now deconstructed thusly:
You outright deny that it is possible for there to be ANY legitimate reason whatsoever for Newsweek to weigh numbers based on information outside YOUR ken. Manifestly absurd.

QED

Oh, BTW, speaking of ignoring the obvious, do you really think that Newsweek pollsters, even with all their lackluster intellectual merit and bias, are too stupid to see the problem you identified by comparing the face value of two numbers?

Maybe, maybe not. Here's what may have happened.

First, a lot of us have cell phones, screening, etc. and never answer the phone. However, while this may vary by party over time, it doesn't week by week. And the people who answer the phone are a different demographic from actual likely voters - the target demographic. Plus, in an automated poll, Junior, Sis, or the repairman can answer the phone, and anyone can lie - like the guy on FR four years ago who said he was a Hispanic Democrat for Bush in 2000 in Pennsylvania. A few like him cause Bush (or Kerry) to mis-allocate their campaign fund and face time.

On to the main point. People who answer the phone and take the poll are, on average, more likely to be weak-minded than are likely voters. Thus, they heard about the convention on TV, and answered the poll that they are voting for Bush. Next week, the Partisan Media Shills (aka PMS) will sway many of these telephone-answerers to say they'll vote for Kerry. Since Nader, Badnarik, etc. don't get as much airtime, they get fewer votes in these polls than at the ballot box. Yes, I know that Wallace in 1968 got 20% in the polls and only 14% in the official vote count. He was also well ahead of Robert Kennedy in Michigan polling in 1968, but lost 20% to 19% in the official count in the DemonRatic primary that year. Some, but not all of the difference can be attributed to voters voting for number 1 or number 2.

IOW, the official numbers are as much suspect as the polls. More so, once you consider that there's much more at stake in the official poll.